1. Lecture 10 Frequent Itemset Mining/Association Rule MW 4:00PM-5:15PM Dr. Jianjun Hu http://mleg.cse.sc.edu/edu/csce822 CSCE822 Data Mining and Warehousing University of South Carolina Department of Computer Science and Engineering

2. 10/16/2015 Roadmap Frequent Itemset Mining Problem Closed itemset, Maximal itemset Apriori Algorithm FP-Growth: itemset mining without candidate generation Association Rule Mining

3. Case 1: D.E.Shaw & Co. D. E. Shaw & Co. is a New York-based investment and technology development firm. By Columbia Uni. CS faculty.  manages approximately US $35 billion in aggregate capital  known for its quantitative investment strategies, particularly statistical arbitragestatistical arbitrage  arbitrage is the practice of taking advantage of a price differential between two or more marketsmarkets  statistical arbitrage is a heavily quantitative and computational approach to equity trading. It involves data mining and statistical methods, as well as automated trading systems

4. StatArb, the trading strategy StatArb evolved out of the simpler pairs trade strategy, in which stocks are put into pairs by fundamental or market-based similarities.pairs trade When one stock in a pair outperforms the other, the poorer performing stock is bought long with the expectation that it will climb towards its outperforming partner, the other is sold short. http://en.wikipedia.org/wiki/Statistical_arbitrage Example: PetroChina SHI CEO

5. StatArb, the trading strategy StatArb considers not pairs of stocks but a portfolio of a hundred or more stocks (some long, some short) that are carefully matched by sector and region to eliminate exposure to beta and other risk factors Q: How can u find those matched/associated stocks? A: Frequent Itemset Mining - Transaction records: S1↑S2↓S3↓S4 ↑ S1↑S2↓S3↑ S4↑ S1↓S2↑S3↓S4 ↓ S1↑S2↓S3↑S4 ↑ (S1, S2)S1 ↓ S2 ↑ Buy S1

6. Case 2: The Market Basket Problem  What products were often purchased together?— Beer and diapers?!  What are the subsequent purchases after buying a PC?  Basket data analysis, cross-marketing, catalog design, sale campaign analysis, Web log (click stream) analysis, and DNA sequence analysis Market-Basket transactions Example of Association Rules {Diaper}  {Beer}, {Milk, Bread}  {Eggs,Coke}, {Beer, Bread}  {Milk}, Implication means co-occurrence, not causality!

7. October 16, 2015 Data Mining: Concepts and Techniques What Is Frequent Pattern Analysis? Frequent pattern: a pattern (a set of items, subsequences, substructures, etc.) that occurs frequently in a data set First proposed by Agrawal, Imielinski, and Swami [AIS93] in the context of frequent itemsets and association rule mining Motivation: Finding inherent regularities in data  What products were often purchased together?— Beer and diapers?!  What are the subsequent purchases after buying a PC?  What kinds of DNA are sensitive to this new drug?  Can we automatically classify web documents? Applications  Basket data analysis, cross-marketing, catalog design, sale campaign analysis, Web log (click stream) analysis, and DNA sequence analysis.

8. October 16, 2015 Data Mining: Concepts and Techniques Why Is Freq. Pattern Mining Important? Discloses an intrinsic and important property of data sets Forms the foundation for many essential data mining tasks  Association, correlation, and causality analysis  Sequential, structural (e.g., sub-graph) patterns  Pattern analysis in spatiotemporal, multimedia, time-series, and stream data  Classification: associative classification  Cluster analysis: frequent pattern-based clustering  Data warehousing: iceberg cube and cube-gradient  Semantic data compression: fascicles  Broad applications

9. Definition: Frequent Itemset Itemset  A collection of one or more items Example: {Milk, Bread, Diaper}  k-itemset An itemset that contains k items Support count (  )  Frequency of occurrence of an itemset  E.g.  ({Milk, Bread,Diaper}) = 2 Support  Fraction of transactions that contain an itemset  E.g. s({Milk, Bread, Diaper}) = 2/5 Frequent Itemset  An itemset whose support is greater than or equal to a minsup threshold

10. Another Format to View the Transaction Data Representation of Database  horizontal vs vertical data layout

11. Closed Patterns and Max-Patterns A long pattern contains a combinatorial number of sub- patterns, e.g., {a 1, …, a 100 } contains ( 100 1 ) + ( 100 2 ) + … + ( 1 1 0 0 0 0 ) = 2 100 – 1 = 1.27*10 30 sub-patterns!  (A, B, C)6 frequent  (A, B) 7, (A, C)6, …also frequent Solution: Mine closed patterns and max-patterns instead  Closed pattern is a lossless compression of freq. patterns  Reducing the # of patterns and rules October 16, 2015 Data Mining: Concepts and Techniques

12. Maximal Frequent Itemset Border Infrequent Itemsets Maximal Itemsets An itemset is maximal frequent if none of its immediate supersets is frequent

13. Closed Itemset An itemset is closed if none of its immediate supersets has the same support as the itemset

14. Maximal vs Closed Itemsets Transaction Ids Not supported by any transactions

15. Maximal vs Closed Frequent Itemsets Minimum support = 2 # Closed = 9 # Maximal = 4 Closed and maximal Closed but not maximal

16. October 16, 2015 Data Mining: Concepts and Techniques Closed Patterns and Max-Patterns Exercise. DB = {, }  Min_sup = 1. What is the set of closed itemset?  : 1  : 2 What is the set of max-pattern?  : 1 What is the set of all patterns?  !!

17. Maximal vs Closed Itemsets

18. Scalable Methods for Mining Frequent Patterns The downward closure property of frequent patterns  Any subset of a frequent itemset must be frequent  If {beer, diaper, nuts} is frequent, so is {beer, diaper}  i.e., every transaction having {beer, diaper, nuts} also contains {beer, diaper} Scalable mining methods: Three major approaches  Apriori (Agrawal & Srikant@VLDB’94)  Freq. pattern growth (FPgrowth—Han, Pei & Yin @SIGMOD’00)  Vertical data format approach (Charm—Zaki & Hsiao @SDM’02) October 16, 2015 Data Mining: Concepts and Techniques 18

19. October 16, 2015 Data Mining: Concepts and Techniques 19 Apriori: A Candidate Generation-and-Test Approach Apriori pruning principle: If there is any itemset which is infrequent, its superset should not be generated/tested! (Agrawal & Srikant @VLDB’94, Mannila, et al. @ KDD’ 94) Method:  Initially, scan DB once to get frequent 1-itemset  Generate length (k+1) candidate itemsets from length k frequent itemsets  Test the candidates against DB  Terminate when no frequent or candidate set can be generated

20. October 16, 2015 20 The Apriori Algorithm—An Example Database TDB 1 st scan C1C1 L1L1 L2L2 C2C2 C2C2 2 nd scan C3C3 L3L3 3 rd scan TidItems 10A, C, D 20B, C, E 30A, B, C, E 40B, E Itemsetsup {A}2 {B}3 {C}3 {D}1 {E}3 Itemsetsup {A}2 {B}3 {C}3 {E}3 Itemset {A, B} {A, C} {A, E} {B, C} {B, E} {C, E} Itemsetsup {A, B}1 {A, C}2 {A, E}1 {B, C}2 {B, E}3 {C, E}2 Itemsetsup {A, C}2 {B, C}2 {B, E}3 {C, E}2 Itemset {B, C, E} Itemsetsup {B, C, E}2 Sup min = 2

21. October 16, 2015 Data Mining: Concepts and Techniques 21 The Apriori Algorithm Pseudo-code: C k : Candidate itemset of size k L k : frequent itemset of size k L 1 = {frequent items}; for (k = 1; L k !=  ; k++) do begin C k+1 = candidates generated from L k ; for each transaction t in database do increment the count of all candidates in C k+1 that are contained in t L k+1 = candidates in C k+1 with min_support end return  k L k ;

22. October 16, 2015 Data Mining: Concepts and Techniques 22 Important Details of Apriori How to generate candidates?  Step 1: self-joining L k  Step 2: pruning How to count supports of candidates? Example of Candidate-generation  L 3 ={abc, abd, acd, ace, bcd}  Self-joining: L 3 *L 3 abcd from abc and abd acde from acd and ace We cannot join ace and bcd –to get 4-itemset  Pruning: acde is removed because ade is not in L 3  C 4 ={abcd}

23. October 16, 2015 Data Mining: Concepts and Techniques 23 How to Generate Candidates? Suppose the items in L k-1 are listed in an order Step 1: self-joining L k-1 insert into C k select p.item 1, p.item 2, …, p.item k-1, q.item k-1 from L k-1 p, L k-1 q where p.item 1 =q.item 1, …, p.item k-2 =q.item k-2, p.item k-1 < q.item k-1 Step 2: pruning forall itemsets c in C k do forall (k-1)-subsets s of c do if (s is not in L k-1 ) then delete c from C k

24. October 16, 2015 Data Mining: Concepts and Techniques 24 How to Count Supports of Candidates? Why counting supports of candidates a problem?  The total number of candidates can be very huge  One transaction may contain many candidates Method:  Candidate itemsets are stored in a hash-tree  Leaf node of hash-tree contains a list of itemsets and counts  Interior node contains a hash table  Subset function: finds all the candidates contained in a transaction

25. October 16, 2015 Data Mining: Concepts and Techniques 25 Example: Store candidate itemsets into Hashtree 1,4,7 2,5,8 3,6,9 Subset function 2 3 4 5 6 7 1 4 5 1 3 6 1 2 4 4 5 7 1 2 5 4 5 8 1 5 9 3 4 5 3 5 6 3 5 7 6 8 9 3 6 7 3 6 8 Hash on z Hash on y For each candidate item (x y z) Hash on first item x

26. First items 1/2/3 October 16, 2015 Data Mining: Concepts and Techniques 26 Example: Counting Supports of Candidates 1,4,7 2,5,8 3,6,9 Subset function 2 3 4 5 6 7 1 4 5 1 3 6 1 2 4 4 5 7 1 2 5 4 5 8 1 5 9 3 4 5 3 5 6 3 5 7 6 8 9 3 6 7 3 6 8 Transaction: 1 2 3 5 6 1 + 2 3 5 6 1 2 + 3 5 6 1 3 + 5 6 5/9 leaf nodes visited 9 out of 15 itemsets compared to transaction 2/3/5 3/5

27. October 16, 2015 Data Mining: Concepts and Techniques 27 Challenges of Frequent Pattern Mining Challenges  Multiple scans of transaction database  Huge number of candidates  Tedious workload of support counting for candidates Improving Apriori: general ideas  Reduce passes of transaction database scans  Shrink number of candidates  Facilitate support counting of candidates

28. October 16, 2015 Data Mining: Concepts and Techniques 28 Bottleneck of Frequent-pattern Mining Multiple database scans are costly Mining long patterns needs many passes of scanning and generates lots of candidates  To find frequent itemset i 1 i 2 …i 100 # of scans: 100 # of Candidates: ( 100 1 ) + ( 100 2 ) + … + ( 1 1 0 0 0 0 ) = 2 100 -1 = 1.27*10 30 ! Bottleneck: candidate-generation-and-test Can we avoid candidate generation?

29. Mining Frequent Patterns Without Candidate Generation Grow long patterns from short ones using local frequent items  “abc” is a frequent pattern  Get all transactions having “abc”: DB|abc  “d” is a local frequent item in DB|abc  abcd is a frequent pattern October 16, 2015

30. FP-growth Algorithm Use a compressed representation of the database using an FP-tree Once an FP-tree has been constructed, it uses a recursive divide-and-conquer approach to mine the frequent itemsets

31. October 16, 2015 Construct FP-tree from a Transaction Database {} f:4c:1 b:1 p:1 b:1c:3 a:3 b:1m:2 p:2m:1 Header Table Item frequency head f4 c4 a3 b3 m3 p3 min_support = 3 TIDItems bought (ordered) frequent items 100{f, a, c, d, g, i, m, p}{f, c, a, m, p} 200{a, b, c, f, l, m, o}{f, c, a, b, m} 300 {b, f, h, j, o, w}{f, b} 400 {b, c, k, s, p}{c, b, p} 500 {a, f, c, e, l, p, m, n}{f, c, a, m, p} 1.Scan DB once, find frequent 1- itemset (single item pattern) 2.Sort frequent items in frequency descending order, f- list 3.Scan DB again, construct FP- tree F-list=f-c-a-b-m-p

32. FP-Tree Construction Example null A:7 B:5 B:3 C:3 D:1 C:1 D:1 C:3 D:1 E:1 Pointers are used to assist frequent itemset generation D:1 E:1 Transaction Database Header table

33. FP-growth null A:7 B:5 B:1 C:1 D:1 C:1 D:1 C:3 D:1 Conditional Pattern base for D: P = {(A:1,B:1,C:1), (A:1,B:1), (A:1,C:1), (A:1), (B:1,C:1)} Recursively apply FP- growth on P Frequent Itemsets found (with sup > 1): AD, BD, CD, ACD, BCD D:1 All transactions that contains the patterns ending with D are encapsulated in this tree.

34. October 16, 2015 Data Mining: Concepts and Techniques 34 Benefits of the FP-tree Structure Completeness  Preserve complete information for frequent pattern mining  Never break a long pattern of any transaction Compactness  Reduce irrelevant info—infrequent items are gone  Items in frequency descending order: the more frequently occurring, the more likely to be shared  Never be larger than the original database (not count node-links and the count field)  For Connect-4 DB, compression ratio could be over 100

35. October 16, 2015 Data Mining: Concepts and Techniques 35 Why Is FP-Growth the Winner? Divide-and-conquer:  decompose both the mining task and DB according to the frequent patterns obtained so far  leads to focused search of smaller databases Other factors  no candidate generation, no candidate test  compressed database: FP-tree structure  no repeated scan of entire database  basic ops—counting local freq items and building sub FP- tree, no pattern search and matching

36. October 16, 2015 Data Mining: Concepts and Techniques 36 Implications of the Methodology Mining closed frequent itemsets and max-patterns  CLOSET (DMKD’00) Mining sequential patterns  FreeSpan (KDD’00), PrefixSpan (ICDE’01) Constraint-based mining of frequent patterns  Convertible constraints (KDD’00, ICDE’01) Computing iceberg data cubes with complex measures  H-tree and H-cubing algorithm (SIGMOD’01)

37. October 16, 2015 Data Mining: Concepts and Techniques 37 MaxMiner: Mining Max-patterns 1 st scan: find frequent items  A, B, C, D, E 2 nd scan: find support for  AB, AC, AD, AE, ABCDE  BC, BD, BE, BCDE  CD, CE, CDE, DE, Since BCDE is a max-pattern, no need to check BCD, BDE, CDE in later scan R. Bayardo. Efficiently mining long patterns from databases. In SIGMOD’98 TidItems 10A,B,C,D,E 20B,C,D,E, 30A,C,D,F Potential max-patterns

38. 10/16/2015 Roadmap Frequent Itemset Mining Problem Closed itemset, Maximal itemset Apriori Algorithm FP-Growth: itemset mining without candidate generation Association Rule Mining

39. Definition: Association Rule Example: l Association Rule –An implication expression of the form X  Y, where X and Y are itemsets –Example: {Milk, Diaper}  {Beer} l Rule Evaluation Metrics –Support (s)  Fraction of transactions that contain both X and Y –Confidence (c)  Measures how often items in Y appear in transactions that contain X

40. Mining Association Rules Example of Rules: {Milk,Diaper}  {Beer} (s=0.4, c=0.67) {Milk,Beer}  {Diaper} (s=0.4, c=1.0) {Diaper,Beer}  {Milk} (s=0.4, c=0.67) {Beer}  {Milk,Diaper} (s=0.4, c=0.67) {Diaper}  {Milk,Beer} (s=0.4, c=0.5) {Milk}  {Diaper,Beer} (s=0.4, c=0.5) Observations: All the above rules are binary partitions of the same itemset: {Milk, Diaper, Beer} Rules originating from the same itemset have identical support but can have different confidence Thus, we may decouple the support and confidence requirements

41. Association Rule Mining Task Given a set of transactions T, the goal of association rule mining is to find all rules having  support ≥ minsup threshold  confidence ≥ minconf threshold Brute-force approach:  List all possible association rules  Compute the support and confidence for each rule  Prune rules that fail the minsup and minconf thresholds  Computationally prohibitive!

42. Mining Association Rules Two-step approach: 1. Frequent Itemset Generation – Generate all itemsets whose support  minsup 2. Rule Generation – Generate high confidence rules from each frequent itemset, where each rule is a binary partitioning of a frequent itemset Frequent itemset generation is still computationally expensive

43. Step 2: Rule Generation Given a frequent itemset L, find all non-empty subsets f  L such that f  L – f satisfies the minimum confidence requirement  If {A,B,C,D} is a frequent itemset, candidate rules: ABC  D, ABD  C, ACD  B, BCD  A, A  BCD,B  ACD,C  ABD, D  ABC AB  CD,AC  BD, AD  BC, BC  AD, BD  AC, CD  AB, If |L| = k, then there are 2 k – 2 candidate association rules (ignoring L   and   L)

44. Rule Generation How to efficiently generate rules from frequent itemsets?  In general, confidence does not have an anti-monotone property c(ABC  D) can be larger or smaller than c(AB  D)  But confidence of rules generated from the same itemset has an anti-monotone property  e.g., L = {A,B,C,D}: c(ABC  D)  c(AB  CD)  c(A  BCD) Confidence is anti-monotone w.r.t. number of items on the RHS of the rule

45. Rule Generation for Apriori Algorithm Lattice of rules Pruned Rules Low Confidence Rule

46. Rule Generation for Apriori Algorithm Candidate rule is generated by merging two rules that share the same prefix in the rule consequent join(CD=>AB,BD=>AC) would produce the candidate rule D => ABC Prune rule D=>ABC if its subset AD=>BC does not have high confidence

47. Pattern Evaluation Association rule algorithms tend to produce too many rules  many of them are uninteresting or redundant  Redundant if {A,B,C}  {D} and {A,B}  {D} have same support & confidence Interestingness measures can be used to prune/rank the derived patterns In the original formulation of association rules, support & confidence are the only measures used

48. Computing Interestingness Measure Given a rule X  Y, information needed to compute rule interestingness can be obtained from a contingency table YY Xf 11 f 10 f 1+ Xf 01 f 00 f o+ f +1 f +0 |T| Contingency table for X  Y f 11 : support of X and Y f 10 : support of X and Y f 01 : support of X and Y f 00 : support of X and Y Used to define various measures u support, confidence, lift, Gini, J-measure, etc.

49. Drawback of Confidence Coffee Tea15520 Tea75580 9010100 Association Rule: Tea  Coffee Confidence= P(Coffee|Tea) = 0.75 but P(Coffee) = 0.9  Although confidence is high, rule is misleading  P(Coffee|Tea) = 0.9375

50. Statistical Independence Population of 1000 students  600 students know how to swim (S)  700 students know how to bike (B)  420 students know how to swim and bike (S,B)  P(S  B) = 420/1000 = 0.42  P(S)  P(B) = 0.6  0.7 = 0.42  P(S  B) = P(S)  P(B) => Statistical independence  P(S  B) > P(S)  P(B) => Positively correlated  P(S  B) Negatively correlated

51. Statistical-based Measures Measures that take into account statistical dependence

52. Example: Lift/Interest Coffee Tea15520 Tea75580 9010100 Association Rule: Tea  Coffee Confidence= P(Coffee|Tea) = 0.75 but P(Coffee) = 0.9  Lift = 0.75/0.9= 0.8333 (< 1, therefore is negatively associated)

53. Drawback of Lift & Interest YY X100 X090 1090100 YY X900 X010 9010100 Statistical independence: If P(X,Y)=P(X)P(Y) => Lift = 1

54. There are lots of measures proposed in the literature Some measures are good for certain applications, but not for others What criteria should we use to determine whether a measure is good or bad? What about Apriori- style support based pruning? How does it affect these measures?

55. Subjective Interestingness Measure Objective measure:  Rank patterns based on statistics computed from data  e.g., 21 measures of association (support, confidence, Laplace, Gini, mutual information, Jaccard, etc). Subjective measure:  Rank patterns according to user’s interpretation A pattern is subjectively interesting if it contradicts the expectation of a user (Silberschatz & Tuzhilin) A pattern is subjectively interesting if it is actionable (Silberschatz & Tuzhilin)

56. Summary Frequent item set mining applications Apriori algorithm FP-growth algorithm Association mining Association Rule evaluation